Dosimetric predictive factors for facial nerve paralysis after cyberknife® stereotactic radiotherapy for vestibular schwannomas: A single institution experience of 88 patients

Abstract

Listen

Translate article

To identify dosimetric predictive factors of facial nerve paralysis for patients with vestibular schwannomas (VS) treated in a single institution with Cyberknife® (CK) hypofractionated stereotactic radiotherapy (SRT). Eighty-eight patients were treated from 2010 to 2020. Different treatment schedules were used over that period, some prescribed to the 80% isodose line (4×5 Gy, 3×7 Gy, 3×8 Gy and 5×5 Gy) and one to the 70% isodose line (3×7.7Gy). Local control tumor and facial nerve toxicity were recorded, as well as various dosimetric indicators. Median follow-up 37months (range, 7-96). Of the 88 stereotactic treatments, 20 patients (23%) developed objectively diagnosed radiation-induced facial nerve paralysis. The 2-year and 5-year local tumor control were respectively 95% and 88%, and the overall 2-year facial nerve preservation was 76%. Prescriptions with a maximum dose point (Dmax) of 33Gy were at a substantially higher risk of facial paralysis than prescriptions with a Dmax less than or equal to 30Gy (HR=4.51, 95% CI=[1.04;19.6], p=0.045). The 2-years cumulative incidences of facial paralysis were 32% [20%;44%] in the case of a 33Gy Dmax, against 7% [1%;21%] otherwise. We identified four significative dosimetric predictive factors for radiation-induced facial nerve dysfunction: a GTV minimal dose over 22Gy (EQD2=45.5Gy, p=0.019), a GTV mean dose over 29Gy (EQD2=73.5Gy, HR=2.84, 95% CI=[1.10;7.36], p=0.024), a PTV mean dose over 27Gy (EQD2=64.8Gy, HR=10.52, 95% CI=[1.39;79.76], p=0.002) and a PTV maximal dose of 32Gy (EQD2=87.5Gy,HR=5.09, 95% CI=[1.17;22.15], p=0.013). We identified four dosimetric predictive factors for post-treatment facial paralysis. Increasing the doses of hypofractionated stereotactic radiotherapy for vestibular schwannomas leads to higher facial nerve toxicity and may lead to lower local control rates than other published series. Our three-hypofractionated regimens may have also played a role in these results.